If we use alot of plug gages and they are noted on our control plans, what type of MSA do you recomend we need to be in compliance with QS?
Thanks, Dawn

Dawn,

I cannot suggest a compliance mechanism, but you can classify plug gages using attribute R&R.

Regards,
Don

[This message has been edited by Don Winton (edited 02-07-99).]

Attribute R & R by what method? Can you explain how I would do this?

I can send by e-mail a Word 6.0 Document that explains R&R for attribute data. Can you read this type of file?

Regards,
Don

Shoot it to me and I'll post it as a pdf file.

On its way. BTW, this is a first draft. If any errors are detected, do let me know, please.

Regards,
Don

I can read from a word file. If Marc posts it that would be great!!!
Thanks, guys!

Question 1: Can anyone recommend a formal way of measuring gage capability on a pass / fail measurement? We have something in-house but I'm not sure what is standardly acceptable. (Our in-house method does not calculate a "gage error", etc) . Question 2: Can anyone recommend a method for an attribute GR&R when the test output is a 1-5 integer ranking based on standards we have set up? Thank you for any suggestions or leads you can provide!

Question 1:
A very good attribute GR&R can be found at Don Winton's website. Perhaps you may find it also in the PDF zone here at the forum. Look for an Excel spreadsheet, it contains most of the directions.

Question 2:
Typically an attribute R&R is simply a pass / fail. If I understand the question, you are creating variables data (albeit not very discriminating) when you classify parts in more than a pass / fail test, therefore a variable R&R should be performed.

Is there anyone out there who can lead me to an answer.
I have simplified the situation in my question.

We reference in our Control Plans to various kinds of Go/No-go Gauges use to inspect significant characteristics. At the end, we have about 5 different types of gauges at our place.

One kind of gauges are the Plug Gauges. I might have one that is for a 10 +/-0.1mm specification and another one that is for a 15 +/-0.1mm specification and 5 others between 1 and 10mm.

Questions:
1) Can Gauge R&R be done for families of gauges (in this instance just for one Plug gauge instead of 7) irrespective of the specification or do we have to do this for each specification?

2) If the same Gauge is being used at 5 totally different parts, can I just refer back to one performed study (criteria specific) instead of doing it 5 times (Part specific)

3) Where would I find those answers in the MSA manual?

Unfortunately there don't seem to be any definitive guidelines about this. Here is my take:
1. Since specs are different, you should not club the gages as one family of 'plug gages'
2. For one gage used on 5 parts, I think you can perform r&r on any one part.
-Atul.

Garry,

I have some experience using plug gages but I'm not an expert, but here's my opinion FWIW: I believe the toughest test for the gage R&R would be for the 15mm +/- 0.1 mm part -- the largest hole size with the tightest tolerance. If the gage is suitable in this application I believe it would only be as good or better for the smaller holes with the same tolerance. I mainly use plug gages for go/no go testing which is easier than using them for quantitiative purposes.

Hope this helps some.

Mike S.

See attached.

Hi
can anyone tell me what is the percentage of fake rejects than an attribute Gauge R&R can have and still be accetable?
After doing one i got 16 fake rejects aout of 1000 products, is this ok?

I have a question regarding a customers demand. They would like for me to cut the tolerance on the part by the amount of error that I see from my Gauge R&R. This is the first that I have heard of this. My understanding of the gauge R&R is to determine if it is the appropiate gauge for the specification to be measured with the tolerances allowed. The customer is suggesting that we do this even if we have less than 10% error. Has anone else ran across this demand or can anyone give an explaination for the reasoning of this?

I have a question regarding a customers demand. They would like for me to cut the tolerance on the part by the amount of error that I see from my Gauge R&R. This is the first that I have heard of this. My understanding of the gauge R&R is to determine if it is the appropiate gauge for the specification to be measured with the tolerances allowed. The customer is suggesting that we do this even if we have less than 10% error. Has anone else ran across this demand or can anyone give an explaination for the reasoning of this?

Tulman

I agree with your understanding. Never heard this one before.

Perhaps you could politely ask them to point out where the MSA manual contains this idea. Or any other authoritative text?

If so, I'd like to read, study and learn from it...you never know!

Sadly you may lose this battle, the customer always wins, even when they are wrong.

I have a question regarding a customers demand. They would like for me to cut the tolerance on the part by the amount of error that I see from my Gauge R&R. This is the first that I have heard of this. My understanding of the gauge R&R is to determine if it is the appropiate gauge for the specification to be measured with the tolerances allowed. The customer is suggesting that we do this even if we have less than 10% error. Has anone else ran across this demand or can anyone give an explaination for the reasoning of this?

This is a fairly common practice called guard-banding. In the normal situation, the producer has a risk of rejecting good product and the customer has a risk of accepting bad product. The extent of this risk is determined by the measurement variation. See MSA for gage performance curves.

Guard-banding places 100% of the risk on the manufacturer and 0% on the customer. Your reject rate may increase, but the customer is protected. I have seen this used frequently when the process capability is much tighter than the tolerances, so the cost of rejects is minimal.

Hi,

I have some questions about attribute MSA.
currently I use Minitab to calculate attribute MSA,
1. use hypothesis test ( p value < 0.05 ) it means the measurement system is accepted
2.also judged by Kappa value( if kappa >0.75, it shows good measurement system)
but if p value < 0.05 and kappa value < 0.75 , what comment should I make ? I can't find the answer in AIAG MSA 3rd.
3. any question is : if currently the result is only go/no-go ,for instances, in visual standard,there are many distinct categories, what method should I use ?
thanks for your reply !!

copperman.

I can send by e-mail a Word 6.0 Document that explains R&R for attribute data. Can you read this type of file?

Regards,
Don


Don,

I am also facing same problem in dealing with attribute data. Can you pls send me the Word 6.0 Document that explains R&R for attribute data to my e-mail: (Edited by Miner. Please contact the poster through Cove PM). I can read this file.